Helllo All,
Need some help understanding parallel processing. I set-up DoParallel and
worked perfectly. I tried to set-up using parallel package following the book
Parallel R but I get the following error:
Error in checkForRemoteErrors(val) :
4 nodes produced errors; first error: 'what' must be a character string or a
function
I have read some posts on the internet suggesting the issue may be related to
S4 class and I have tried parse() to no avail. I have read Snow and Parallel
documentation but I am not getting the issue.
Glenn
library(parallel)
library(foreach)
library(doParallel)
From Parallel R and Internet set up the following
============ DoParallel 4 seconds (perfect) took about 10 mins to set-up
==============
ptm <- proc.time()
Cluster <- makeCluster(detectCores())
registerDoParallel(Cluster)
clusterSetRNGStream(cl = Cluster, set.seed(100))
OAS <- foreach(i = 1:4, .packages = c("BondLab"), .combine = cbind) %dopar%
Mortgage.OAS(bond.id = "bondlabMBS4", trade.date = "01-10-2013", settlement.date =
"01-13-2013",
original.bal = 100000, price = 105.75, sigma = .23/sqrt(240), paths = 50)@OAS
mean(OAS)
proc.time() - ptm
hist(OAS * 100, breaks = 20)
stopCluster(Cluster)
============== followed the book example fails with the following message:
Error in checkForRemoteErrors(val) :
4 nodes produced errors; first error: 'what' must be a character string or a
function ============
ptm <- proc.time()
RNGkind("L'Ecuyer-CMRG")
cl <- makeCluster(detectCores())
clusterSetRNGStream(cl, set.seed(100))
clusterEvalQ(cl, library(BondLab))
OAS <- clusterApply(cl = cl, 1:4, Mortgage.OAS(bond.id = "bondlabMBS4",
trade.date = "01-10-2013", settlement.date = "01-13-2013",
original.bal = 100000, price = 105.75, sigma = .23/sqrt(240), paths =
50)@OAS)
proc.time() - ptm
mean(OAS)
hist(OAS * 100, breaks = 20)
stopCluster(cl)
______________________________________________
R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
